Machine Learning for the Discovery, Design, and Engineering of Materials.

Area of Science:

• Materials Science
• Computational Chemistry
• Artificial Intelligence

Background:

• Machine learning (ML) is integral to high-throughput screening and computational materials discovery.
• Challenges persist in fully leveraging ML for robust materials engineering and advanced design strategies.

Purpose of the Study:

• To review recent advances in ML algorithms and their applications in materials science.
• To highlight ML's role in accelerating materials discovery, design, and engineering.
• To identify future opportunities for ML in computational materials design.

Main Methods:

• Review of recent advancements in ML algorithms and their applications.
• Analysis of ML's integration with physics-based models for improved performance.
• Examination of ML's utility in large-scale screening, rule identification, and multi-objective engineering.

Main Results:

• ML can outperform, accelerate, or integrate with physics-based models depending on data availability.
• ML facilitates the discovery of new materials via large-scale screening.
• ML aids in designing materials by uncovering governing principles and engineering for multiple objectives.

Conclusions:

• ML is making significant inroads in materials discovery, design, and engineering.
• Further advancements are needed to establish ML as a standard tool for practical computational materials design.
• ML offers a promising pathway beyond traditional trial-and-error and screening methods.